highly significantp < 0.001
First, hypothesis-driven statistical filtering retained only variables demonstrating highly significant intergroup differences ( p < 0.001) between DLC and NDLC cohorts, thereby substantially reducing the risk of Type I error.